Balancer Test Results!!

General Comments, Questions about all things OmnipotentS that don't go in other topics/forums
User avatar
pooty
Posts: 4358
Joined: Sat Apr 03, 2021 10:22 am
Location: Michigan
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by pooty »

Tater wrote: Tue Jul 06, 2021 2:22 pm I agree it does create a toxic learning curve for many newbies, which is really the reason we see some names on the server once and never again. The server (and sometimes certain people) are very non-friendly to new players.
Yes, we should be patient, and friendly with new players...and not be too hard on them, even if they take a "power" vehicle and noob it. Heck, they probably wonder why the Mino tank drives so badly....
And most Admins can look to see if its a new player and not someone Aliasing. I usually check before I warn or kick to make sure they aren't new.... and some players like Gunshy, like to constantly Alias, but he talks and you usually know him right away.
User avatar
pooty
Posts: 4358
Joined: Sat Apr 03, 2021 10:22 am
Location: Michigan
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by pooty »

So the topic came up that a 4-0 and 3-0 victory are almost the same thing...I might even argue that a 4-1 is more or less a good ass whooping.
If you combine the numbers:
Say 8v8:
3 point or more victory, 2 pt victory (4-2, 3-1), 1pt
ON: 42, 60%, 13 18%, 14 19%
OFF: 75, 70%, 21 20%, 10, 10%

So definitely more ass whoopings eg. lopsided victories with Balancer OFF; again look at the percentages. The number of close matches DOUBLED (10% to 19%) and lopsided victories was reduced by 14% Did the Balancer fix lopsided victories? Somewhat, but what isn't shown is even in a 3-0 match, which might have been 3 1pt rounds, all going to OT -- could even have won by just a few core points each time -- gets counted the same as a fast 2pt round, followed by a 1pt round...and those aren't the same type of match.
User avatar
Enyo
Posts: 1626
Joined: Mon Apr 05, 2021 11:27 pm
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by Enyo »

pooty wrote: Wed Jul 07, 2021 8:40 am So the topic came up that a 4-0 and 3-0 victory are almost the same thing...I might even argue that a 4-1 is more or less a good ass whooping.
If you combine the numbers:
Say 8v8:
3 point or more victory, 2 pt victory (4-2, 3-1), 1pt
ON: 42, 60%, 13 18%, 14 19%
OFF: 75, 70%, 21 20%, 10, 10%

So definitely more ass whoopings eg. lopsided victories with Balancer OFF; again look at the percentages. The number of close matches DOUBLED (10% to 19%) and lopsided victories was reduced by 14% Did the Balancer fix lopsided victories? Somewhat, but what isn't shown is even in a 3-0 match, which might have been 3 1pt rounds, all going to OT -- could even have won by just a few core points each time -- gets counted the same as a fast 2pt round, followed by a 1pt round...and those aren't the same type of match.
Maybe it would be worth looking at the time logged for each of those matches to see how short/long they were. If there's a 4-0 or 3-0 victory, if the entire match time was shorter than say 15 minutes, it was a quick blowout, but if the time is 25+ minutes, it obviously took time for the winning team to wear down the other. Or, it got to a choke point the winning team couldn't break through, or maybe some players left/joined. I don't know what the total match time threshold should be for considering the blowout victories, but lumping in 30-40 minute matches with the "blowouts" results in bad data IMO.

The main problem I have with the balancer is the way it mixes teams up. I feel like it constantly takes one or two top-scoring players and pairs them with a bunch of shlubs, then pairs a couple slightly less top-scoring players with a bunch of mid-level players... the latter team steamrolls the former every time. And when you have roughly the same set of 20 players, it seems to divide them up almost the same way every time, maybe changing one or 2 players. That is a problem that needs to be addressed if the balancer is to be left on.
“Never argue with stupid people, they will drag you down to their level and then beat you with experience.”
― Mark Twain
User avatar
captainsnarf
Posts: 2632
Joined: Tue Jul 06, 2021 1:51 pm
Location: Washington
Server Sponsor: Yes
Server Admin: Yes
Contact:

Re: Balancer Test Results!!

Post by captainsnarf »

I have a really-simple-maybe-too-obvious solution to this.

Don't use stats.

There aren't a lot of people playing anymore. If you list the unique players that join the server, it's probably < 50 people. Take the list of 50 people and rank them manually yourself. Any player that joins not in the list can go back to the old stats method.

Players like Tator, Enyo, Leon, Xexx will be at the top. They should not all be on the same team :)
User avatar
YEAAAHHHHHHHHHH
Posts: 1058
Joined: Tue Jun 08, 2021 3:03 pm
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by YEAAAHHHHHHHHHH »

Too bad you can't play with "playground rules". Each team gets a captain and you pick your team one at a time before each match.
User avatar
McLovin
Posts: 1165
Joined: Sat Apr 03, 2021 12:54 pm
Location: Salt Lake City, Utah
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by McLovin »

captainsnarf wrote: Wed Jul 07, 2021 7:42 pm ...
and rank them manually yourself.
...
I suggested this, but nobody felt bold enough to do the ranking. Are you volunteering? Anybody... Bueller...

I don't think it would be that hard for somebody that plays regularly and is familiar with everybody and their playing skills. I say do it and give up trying to qualify/quantify a moving target with statistics.

BTW: This is exactly how they do it at my workplace when a RIF (Reduction In Force) is about to happen. They rank everybody from top to bottom, most valuable (can't do without), to least useful (get rid of them at the next opportunity). Then HR (Human Resources) goes through the list and maps their salary, total benefit package, age brackets, etc, and creates a top-to-bottom list. They can't be held accountable for any plausible discrimination, so the company lawyers take a look. Also they cut the bottom 5% every year and add 5% new hires to make up for them, that is standard operating procedure for business.
User avatar
McLovin
Posts: 1165
Joined: Sat Apr 03, 2021 12:54 pm
Location: Salt Lake City, Utah
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by McLovin »

YEAAAHHHHHHHHHH wrote: Wed Jul 07, 2021 11:51 pm Too bad you can't play with "playground rules". Each team gets a captain and you pick your team one at a time before each match.
It would be fun to do this, clan-style? Maybe a special event with fixed teams, nobody enters or leaves, but agrees to play out the entire match.

The biggest problem with implementing a good balancer is you only get one shot at the beginning of a match. Nobody wants to get switched mid-match, ie, Mailbox-mulligan-style. With players constantly joining, leaving, spectating, etc the team dynamics is in flux the moment somebody joins/leaves.
User avatar
pooty
Posts: 4358
Joined: Sat Apr 03, 2021 10:22 am
Location: Michigan
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by pooty »

So the balancer has been OFF for a few days, and I haven't played as much as I'd like but I've gotten some feedback that quite often the teams were pretty bad. Experienced that myself two days ago. For all it faults, and lack of perfection, I think it'd be better on.
While at times I think the balancer penalizes a few of the top tier players: It generally puts a very high player first, then uses points to balance it out.. So you might get a 500pt player and 200 pt and 100 pt player vs. 350, 300, 200 on a team based game even a super capable player can't outwork 3 decent players...

I still want to figure out how to let the admin force a rebalance (points based) shuffle to account for skewed teams/joins/leaves.
User avatar
McLovin
Posts: 1165
Joined: Sat Apr 03, 2021 12:54 pm
Location: Salt Lake City, Utah
Server Sponsor: Yes
Server Admin: Yes

Re: Balancer Test Results!!

Post by McLovin »

pooty wrote: Thu Jul 15, 2021 1:39 pm ... I think it'd be better on.
I agree.
pooty wrote: Thu Jul 15, 2021 1:39 pm I still want to figure out how to let the admin force a rebalance (points based) shuffle to account for skewed teams/joins/leaves.
Is there an admin command to manually switch players from one team to another? I might have already asked this before but I forgot about it because I've never done it.
User avatar
Tater
Posts: 114
Joined: Wed Jun 16, 2021 12:11 am

Re: Balancer Test Results!!

Post by Tater »

pooty wrote: Thu Jul 15, 2021 1:39 pm I think it'd be better on.

I still want to figure out how to let the admin force a rebalance (points based) shuffle to account for skewed teams/joins/leaves.
Agree. I started playing again last night and the teams were ridiculous. We had a game (I think on mtmu) where everyone on my side was 30+ points, meanwhile the otherside's top player had 15.
Not perfect, but better than literal random teams. The data backs this up.
Some make the claim that the admins can swap around to even the teams, but I can't recall that happening more than *maybe* once.
Post Reply